The worst programmers I know started with C. They write code with ad-hoc, prematurely optimized and unreuseable data structures that is riddled with memory leaks and completely unmaintainable. Then they call you a wuss when you want to write something in a language which enforces a modicum of type safety and provides garbage collection and language constructs like polymorphism which encourage code reuse and abstraction of detail. I don't blame them so much as I blame the amateur hackers who introduced them to it who have no regard or understanding of what responsible software engineering entails. I usually find that it is a better use of my time to berate these people until they start making unbiased choices with regard to their choice of tools and methods than trying to decipher their gibberish, though sometimes deadlines give me no choice. There are plenty of good uses for C, but there are lots of applications where more powerful languages save a lot of pain.
Check out computational neurobiology
on
AI Going Nowhere?
·
· Score: 3, Interesting
Actually, they have made significant strides already in figuring out how the brain works. Check out the Levy Lab at the University of Virginia. They've trained a computer model of a rat's hippocampus to do all sorts of intelligent things, such as transitive inference, sequence completion/combination/disambiguatuion, goal finding, etc. While these are not difficult problems for humans to solve or hardcode into a program, the fact that a single network can do these different and sometimes contradictory things represents something that I would call intelligence. As far as I know, they don't plan on having a model of a human brain very soon, since U.Va. lacks NSA-scale compute servers, but even rat-level learning is pretty cool.
...that inconvenience makes any system less secure, because lazy people will do stupid things to alleviate the inconvenience. This seems like a step in the wrong direction.
Actually, ALL of the teams make their dogs crawl on the elbows, and have for a while. They can actually move them faster that way, because they're more stable. Shielding the ball while still moving with it effectively is much more difficult, and not everyone has that part figured out.
MS gave the MSUG at my school 100 similarly marked CDs for XP Pro. While I'm not in MSUG, I went to one of their meetings, and they gave me a copy. Half the CS department has them. Anyone else have similar experiences at their schools?
Kurdish territory already has some cell phone service. I'm gonna take a shot in the dark and guess that it's the same that's being used everywhere else.
I used to think the imperialistic aspects of this war were secondary to the ideological ones. Lately my opinion has been headed in the other direction, not that either one is good.
From what I've heard, it doesn't sound like he broke any security. He just wrote a script that attempted logins with various SSNs, and recorded successes. The site login was so braindead that it didn't even cross-reference against last name. Unlike a dictionary attack to guess a password, which is given the legal presumption of being private data, he was tossing up perfectly public SSNs. He's apparently also being charged with impersonating another in the commission of a felony, or something like that. Depending on what the site login screen asks, he may have been impersonating other people, but if he's not gaining access improperly, because he's only using public data anyway, then that doesn't apply because he wasn't committing a felony to begin with.
Waves have characteristic "shapes" in both frequency and time domains depending on what created them. A pulse of waves would itself be a wave, and if the output from a pulsar is not a simple sinusoidal wave, it's probably some other form from which a distinct start and end can be determined. There are some forms that come to mind that do not lend themselves to a meaningful determination of start and finish, like exponentials, but those would have infinite energy output anyway, so they're not too likely. The computer simulations they did support these results, and we all know that simulations are always right. So either they've found something new and interesting, or they've found something not so new and interesting that will require them to refine their model of how pulsars work anyway. Either way, something interesting has been learned.
I saw Boneh give a talk on this a few days ago. One machine, over a LAN, can do it in two hours. Sure, that can be noticed if you know to look for lots of aborted SSL sessions, but if you don't have this on an IDS or something like that, someone can Nimda your secretary's box and still get your private key before your sysadmin gets back from the weekly staff meeting.
You'd have to be nuts to take this deal. Microsoft is a gigantic corporation with an extremely diverse product portfolio. This deal would commit you for 10 years (an eternity in the software business) to reject them without consideration. Perhaps Microsoft is the only company to have a product with feature X that your students and professors need. Aside from the fact that this is quite likely illegal, to force an entire school to not use some system, which you'd effectively be doing since any software the students are expected to use must be installed on lab computers or otherwise be made available, goes completely against the concept of intellectual freedom that academia holds so dear.
The students don't get any benefit by being forced to go non-microsoft. They get benefit from being shown all the options. They're going to go out into the world in 4 years and damn near every last one of them is gonna need at least some familiarity with MS products. How much exposure they get should be up to them.
I'm a student at a university where PCs, Macs, and Unix are widely distributed, and I've found that to be a great advantage to me. I don't use MS at all on my personal system, but an awful lot of people do. I work tech support at my university and support the mostly MS userbase. I write software to run on Linux and document it in Word, because I'm familiarizing myself with all of the available tools of my chosen trade.
At my university, those few privileged enough to have unix machines in offices under their complete control often set up IPsec tunnels from their dorm machines, because the dorm net connection to the outside world is prioritized, and anything other than port 80, port 22, port 5190 (AIM) and a few others goes painfully slow. The main campus network is not on a prioritized router, and the connection between the dorm net and the main net is not either, so people use that to play nice low-latency quake. If they implement something like that here, unless they're metering at the switch to the dorm room, people will get around it.
I work in computing support for an academic institution which shall remain nameless. My observation has been that we are generally more secure than most other academic institutions. That being said, I once helped someone who was trying to transfer a rather large file from a satellite office to one of our main offices. The person had been having problems with an FTP server. I checked the server in question. Lo and behold, there was a text file with 50 screens worth of SSNs, names, and addresses, on a publically readable server.
Academic institutions frequently fall victim to the security/convenience tradeoff. While the official policy may be to err on the side of security, an awful lot of people with access to sensitive data don't have any data security training at all, and just "do what works". With a large bureaucracy, the odds that at least one person will screw up are rather high.
Once again, the extreme depth of our knowledge about semiconductors, as a result of their usefulness in computers over the past few decades, lends itself to a rather novel and unrelated use.
My brother and I were working on our 486, and knocked almost 12 oz. of ginger ale into a running 486, hitting all of the cards and part of the motherboard. For reasons unclear, the machine kept operating correctly (even on Win95!), though we did shut it down and clean everything off for safety. No damage at all. We're still using most of those components in newer machines now.
Actually, AI researchers now make the distinction between systems that think/act rationally, and systems that think/act like humans. Either way it's AI. After decades of watching the results of their experiments, they decided that using humans as a sole comparison was pretty messed up, because humans themselves are pretty messed up. AI gives us the chance to create something that is in some ways more than human. Expecting all of them to be bug-for-bug compatible with us is an understandable mistake for those who made it decades ago, but now that we've seen the amazing things AI can do, we should know better.
I always preferred to think of it as a nipple, and I found it to be MUCH easier to use than those damn trackpads. I had to get a cheap hp laptop with a trackpad for economic reasons, but I really miss that thinkpad nipple. Or real nipples. I miss those too. Dammit.
There's not really much we can do about an asteroid by the time we know whether or not it's headed for a populated area. The NASA project is only looking for extinction-level asteroids, not city-killers, because there are just too many of those. We would hope to find a way to push a planet-killer off course if given enough warning, but with a city-killer, you don't know where it's going to hit until it's far too close to do anything about it.
I once programmed a robot that had 16k of flash for basic code (it had an on-board interpreter), and 32 BYTES of memory. Since it was only doing one task, we were able to actually give it some learning ability, and do it only using 11 bytes of memory. We used a good chunk of the flash instruction space as well. While it is not hard to make a simple system very compact, they fact that the Atari developers were able to force an inherently complicated system like a video game to be sufficiently simple that it would fit into rather limited space is truly impressive.
In the U.S., the key to double jeopardy is that they cannot try you for the same crime, but they can try you again for the same act. Crimes have elements. If the elements of one crime make a subset of the elements of a second crime, that first crime is a lesser included offense of the second. You cannot be tried again on a lesser included offense because you were implicitly tried for it already in the first trial. If the government wants to try you for another crime that has at least one element that was not in the elements for the crime you were acquitted of, they are allowed to do that.
Usually, the relevant state and federal laws are worded sufficiently differently that the elements of the crime are not a perfect match, so they can try you again for essentially the same act.
Anyone know what Norwegian law has to say about double jeopardy?
The range is a bit wider than 7-17%. The amorphous silicon cells typically used on houses and in cheap devices like handheld calculators are typically around 5% or lower. In contrast, during the 2001 American Solar Challenge, there were two teams with space-grade Gallium Arsenide arrays, at 24% and 26% efficiency. Interestingly enough, they were small Christian colleges who apparently knew the right people. The winning University of Michigan team, which apparently had a million dollar budget, and exceeded it, couldn't even afford something that high grade. So there is higher stuff out there. It's expensive as hell though.
On a side note, the second place University of Missouri at Rolla team only had a 14% efficient array, which places their power output near the bottom of the open class. The lesson here is that we could save an awful lot of money by learning to do more with the energy we already have available to us.
Slashdot Mirror
The worst programmers I know started with C. They write code with ad-hoc, prematurely optimized and unreuseable data structures that is riddled with memory leaks and completely unmaintainable. Then they call you a wuss when you want to write something in a language which enforces a modicum of type safety and provides garbage collection and language constructs like polymorphism which encourage code reuse and abstraction of detail. I don't blame them so much as I blame the amateur hackers who introduced them to it who have no regard or understanding of what responsible software engineering entails. I usually find that it is a better use of my time to berate these people until they start making unbiased choices with regard to their choice of tools and methods than trying to decipher their gibberish, though sometimes deadlines give me no choice. There are plenty of good uses for C, but there are lots of applications where more powerful languages save a lot of pain.
Actually, they have made significant strides already in figuring out how the brain works. Check out the Levy Lab at the University of Virginia. They've trained a computer model of a rat's hippocampus to do all sorts of intelligent things, such as transitive inference, sequence completion/combination/disambiguatuion, goal finding, etc. While these are not difficult problems for humans to solve or hardcode into a program, the fact that a single network can do these different and sometimes contradictory things represents something that I would call intelligence. As far as I know, they don't plan on having a model of a human brain very soon, since U.Va. lacks NSA-scale compute servers, but even rat-level learning is pretty cool.
Will it play the audio on the trailer for The Matrix Reloaded?
...that inconvenience makes any system less secure, because lazy people will do stupid things to alleviate the inconvenience. This seems like a step in the wrong direction.
Actually, ALL of the teams make their dogs crawl on the elbows, and have for a while. They can actually move them faster that way, because they're more stable. Shielding the ball while still moving with it effectively is much more difficult, and not everyone has that part figured out.
I was about to express my pity for the poor guy whose overloaded server is hosting the article, when I realized that he's the one who slashdotted it!
MS gave the MSUG at my school 100 similarly marked CDs for XP Pro. While I'm not in MSUG, I went to one of their meetings, and they gave me a copy. Half the CS department has them. Anyone else have similar experiences at their schools?
Pretty soon, the car will be free too, as long as you purchase 100,000 gallons of gas.
Kurdish territory already has some cell phone service. I'm gonna take a shot in the dark and guess that it's the same that's being used everywhere else.
I used to think the imperialistic aspects of this war were secondary to the ideological ones. Lately my opinion has been headed in the other direction, not that either one is good.
From what I've heard, it doesn't sound like he broke any security. He just wrote a script that attempted logins with various SSNs, and recorded successes. The site login was so braindead that it didn't even cross-reference against last name. Unlike a dictionary attack to guess a password, which is given the legal presumption of being private data, he was tossing up perfectly public SSNs. He's apparently also being charged with impersonating another in the commission of a felony, or something like that. Depending on what the site login screen asks, he may have been impersonating other people, but if he's not gaining access improperly, because he's only using public data anyway, then that doesn't apply because he wasn't committing a felony to begin with.
Waves have characteristic "shapes" in both frequency and time domains depending on what created them. A pulse of waves would itself be a wave, and if the output from a pulsar is not a simple sinusoidal wave, it's probably some other form from which a distinct start and end can be determined. There are some forms that come to mind that do not lend themselves to a meaningful determination of start and finish, like exponentials, but those would have infinite energy output anyway, so they're not too likely. The computer simulations they did support these results, and we all know that simulations are always right. So either they've found something new and interesting, or they've found something not so new and interesting that will require them to refine their model of how pulsars work anyway. Either way, something interesting has been learned.
I saw Boneh give a talk on this a few days ago. One machine, over a LAN, can do it in two hours. Sure, that can be noticed if you know to look for lots of aborted SSL sessions, but if you don't have this on an IDS or something like that, someone can Nimda your secretary's box and still get your private key before your sysadmin gets back from the weekly staff meeting.
You'd have to be nuts to take this deal. Microsoft is a gigantic corporation with an extremely diverse product portfolio. This deal would commit you for 10 years (an eternity in the software business) to reject them without consideration. Perhaps Microsoft is the only company to have a product with feature X that your students and professors need. Aside from the fact that this is quite likely illegal, to force an entire school to not use some system, which you'd effectively be doing since any software the students are expected to use must be installed on lab computers or otherwise be made available, goes completely against the concept of intellectual freedom that academia holds so dear.
The students don't get any benefit by being forced to go non-microsoft. They get benefit from being shown all the options. They're going to go out into the world in 4 years and damn near every last one of them is gonna need at least some familiarity with MS products. How much exposure they get should be up to them.
I'm a student at a university where PCs, Macs, and Unix are widely distributed, and I've found that to be a great advantage to me. I don't use MS at all on my personal system, but an awful lot of people do. I work tech support at my university and support the mostly MS userbase. I write software to run on Linux and document it in Word, because I'm familiarizing myself with all of the available tools of my chosen trade.
At my university, those few privileged enough to have unix machines in offices under their complete control often set up IPsec tunnels from their dorm machines, because the dorm net connection to the outside world is prioritized, and anything other than port 80, port 22, port 5190 (AIM) and a few others goes painfully slow. The main campus network is not on a prioritized router, and the connection between the dorm net and the main net is not either, so people use that to play nice low-latency quake. If they implement something like that here, unless they're metering at the switch to the dorm room, people will get around it.
I work in computing support for an academic institution which shall remain nameless. My observation has been that we are generally more secure than most other academic institutions. That being said, I once helped someone who was trying to transfer a rather large file from a satellite office to one of our main offices. The person had been having problems with an FTP server. I checked the server in question. Lo and behold, there was a text file with 50 screens worth of SSNs, names, and addresses, on a publically readable server.
Academic institutions frequently fall victim to the security/convenience tradeoff. While the official policy may be to err on the side of security, an awful lot of people with access to sensitive data don't have any data security training at all, and just "do what works". With a large bureaucracy, the odds that at least one person will screw up are rather high.
Once again, the extreme depth of our knowledge about semiconductors, as a result of their usefulness in computers over the past few decades, lends itself to a rather novel and unrelated use.
Cool!
My brother and I were working on our 486, and knocked almost 12 oz. of ginger ale into a running 486, hitting all of the cards and part of the motherboard. For reasons unclear, the machine kept operating correctly (even on Win95!), though we did shut it down and clean everything off for safety. No damage at all. We're still using most of those components in newer machines now.
Actually, AI researchers now make the distinction between systems that think/act rationally, and systems that think/act like humans. Either way it's AI. After decades of watching the results of their experiments, they decided that using humans as a sole comparison was pretty messed up, because humans themselves are pretty messed up. AI gives us the chance to create something that is in some ways more than human. Expecting all of them to be bug-for-bug compatible with us is an understandable mistake for those who made it decades ago, but now that we've seen the amazing things AI can do, we should know better.
Do you think new laws that allow ISPs and end-users to collect damages from spammers on a per-message basis can be effective tools to reduce spam?
I always preferred to think of it as a nipple, and I found it to be MUCH easier to use than those damn trackpads. I had to get a cheap hp laptop with a trackpad for economic reasons, but I really miss that thinkpad nipple. Or real nipples. I miss those too. Dammit.
There's not really much we can do about an asteroid by the time we know whether or not it's headed for a populated area. The NASA project is only looking for extinction-level asteroids, not city-killers, because there are just too many of those. We would hope to find a way to push a planet-killer off course if given enough warning, but with a city-killer, you don't know where it's going to hit until it's far too close to do anything about it.
I once programmed a robot that had 16k of flash for basic code (it had an on-board interpreter), and 32 BYTES of memory. Since it was only doing one task, we were able to actually give it some learning ability, and do it only using 11 bytes of memory. We used a good chunk of the flash instruction space as well. While it is not hard to make a simple system very compact, they fact that the Atari developers were able to force an inherently complicated system like a video game to be sufficiently simple that it would fit into rather limited space is truly impressive.
In the U.S., the key to double jeopardy is that they cannot try you for the same crime, but they can try you again for the same act. Crimes have elements. If the elements of one crime make a subset of the elements of a second crime, that first crime is a lesser included offense of the second. You cannot be tried again on a lesser included offense because you were implicitly tried for it already in the first trial. If the government wants to try you for another crime that has at least one element that was not in the elements for the crime you were acquitted of, they are allowed to do that.
Usually, the relevant state and federal laws are worded sufficiently differently that the elements of the crime are not a perfect match, so they can try you again for essentially the same act.
Anyone know what Norwegian law has to say about double jeopardy?
The range is a bit wider than 7-17%. The amorphous silicon cells typically used on houses and in cheap devices like handheld calculators are typically around 5% or lower. In contrast, during the 2001 American Solar Challenge, there were two teams with space-grade Gallium Arsenide arrays, at 24% and 26% efficiency. Interestingly enough, they were small Christian colleges who apparently knew the right people. The winning University of Michigan team, which apparently had a million dollar budget, and exceeded it, couldn't even afford something that high grade. So there is higher stuff out there. It's expensive as hell though.
On a side note, the second place University of Missouri at Rolla team only had a 14% efficient array, which places their power output near the bottom of the open class. The lesson here is that we could save an awful lot of money by learning to do more with the energy we already have available to us.
The Xbox2 will have 3 Teraflops peak performance, will cost a lot more, and will have a vast array of forgettably unprofitable games.